On Tue, 6 Feb 2018 10:49:38 -0800, Tom Ross <tmr...@stlvm20.vnet.ibm.com> wrote:

>IBM COBOL development needs your help!
>
>We are reviewing a request to change our support for OPTFILE and SYSOPTF
>to allow usage of DD SYSOPTF without the compiler option OPTFILE.
>
>For background, this is where you can avoid the 255 character limit for
>PARM= in JCL when specifying COBOL compiler options.  Currently, if you
>specify compiler option OPTFILE, the compiler tries to OPEN the file
>allocated to DD SYSOPTF, and read compiler options from that file.
>
>OK, we got an RFE (Request For Enhancement) to have the compiler always
>try to use SYSOPTF, with or without the OPTFILE compiler option.  The use
>of SYSOPTF would then only be controlled by the existence of SYSOPTF.
>
>Our concern is, would this affect current users of SYSOPTF?  Are there users
>of SYSOPTF with COBOL who sometimes compile with NOOPTFILE and leave the
>DD statement for SYSOPFT in their JCL/Changeman compile jobs?
>If so, then automatically accessing SYSOPTF without using OPTFILE could
>cause problems.
>
>This leads to another question...do any of your shops  use OPTFILE and
>SYSOPTF for COBOL compiles?
>
>Cheers,
>TomR              >> COBOL is the Language of the Future! <<

We use OPTFILE to supply Compiler defaults.  We specify it as the first parm to 
the Compiler and then specify any options we want to override or options that 
are not in the defaults after OPTFILE.  Since we always use OPTFILE, we always 
have a SYSOPTF DD statement supplied also so we would be OK with this change.  
However, I see the following statements in the COBOL Programming Guide:

The precedence of options in the SYSOPTF data set is determined by where you
specify the OPTFILE option. For example, if you specify OPTFILE in the 
invocation
PARM string, an option specified later in the PARM string supersedes any option
specified in the SYSOPTF data set that conflicts with it.

(Conceptually, OPTFILE in an options specification is replaced with the options 
that
are in the SYSOPTF data set; then the usual rules about precedence of compiler
options and conflicting compiler options apply.)


This seems to imply that options specified as Compiler Parms before OPTFILE 
could be overridden by options in SYSOPTF.  If that's the case and COBOL was 
changed to always use SYSOPTF, would the options in SYSOPTF be processed before 
any Parm options or after all Parm options?

I see in your latest posting that you will probably need to add a new Compiler 
option.  The manual also says that OPTFILE cannot be set as an installation 
default option.  Maybe if it was allowed to be an installation default, that 
would satisfy the RFE?  Maybe with an additional parameter for an installation 
default to specify when SYSOPTF is to processed, first or last?

-- 
Dale R. Smith   

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to